Automatic system architecture synthesis for FPGA-based reconfigurable computers

The goal of this PhD project is to develop an automatic method of system architecture synthesis for general high-performance applications on FPGA-based reconfigurable computers. Through our previous research, we have built a theoretical model targeting the scheduling problem with first-order hardware constraints. And a list scheduling algorithm is developed to achieve near-optimal performances. Currently, we are working on the low-level implementation. A systolic architecture is used, and the list scheduling algorithm will be extended to take into account constraints deriving from exact hardware architecture.